Okay... That's a baller move.
we're changing tpuf's storage architecture so we can keep pushing the search frontier
the new engine (v3) now passes CI, but we have to make it fast before we ship it to prod
we're grinding query plans in public until v3 wins on everything
follow along: turbopuffer.com/v3
Michael Bang retweeted
New @CMUDB Seminar Series: Whatever Andy Found Interesting While Out At Sea
An eclectic collection of database technical talks. Mondays @ 4:30pm ET starting Sep 21 over Zoom. Youtube afterwards. db.cs.cmu.edu/seminars/fall2…
ALT Whatever Andy Found Interesting While Out At Sea — Fall 2026 https://db.cs.cmu.edu/seminars/fall2026/
Michael Bang retweeted
It took myself and another engineer three weeks to build the VOPR.
I mean, what’s the cost of never investing in determinism and in your own simulator?
And which is going to give you that 100x bug finding (and fixing) velocity?
We get so far ahead on TigerStyle, honestly anything else is 1% marginal.
Michael Bang retweeted
Do you like databases?
Do you want to hear two database professors rant about them?
Do you need one of those professors to have a Turing Award for databases?
If yes, then join Mike Stonebraker and I next Wed Dec 10 @ 1:00pm EST for database hot takes:
dbos.dev/webcast-2025-in-rev…
Okay, this is a first 🙈
I'm hoping to find the time to write a series of posts over the summer on tools that I love. Here's the first one which I fell in love with just 3 months ago: mise
blog.vbang.dk/2025/06/29/too…
I'm hoping to find the time to write a series of posts over the summer on tools that I love. Here's the first one which I fell in love with just 3 months ago: mise
blog.vbang.dk/2025/06/29/too…
I've been like yeah, yeah, Thorsten, amp sure does sound nice. But come on.. I've got Claude in VSCode agent mode.
Amp is a _whole_ different game. First try, it fixed two large problems I've had with a Vue frontend app that I've dreaded having to fix. 1/2
God, I love @jdxcode's mise. I'm reading @thorstenball's newsletter which convinced me to try jujutsu (again), and thought to myself "you got me, let's try again"
Just `mise use jujutsu` later and it's there, ready for me to play around with. Why isn't more software like this?
Tired of software relicensing and BSL, but don't know how to navigate it? @jorandirkgreef has a very interesting take on it. youtube.com/watch?v=Z0kzYlCT…
This is preaching to the choir, but I'm sure there's a great podcast episode here, nudge nudge @bcantrill @ahl
Michael Bang retweeted
Today @peer_rich is introducing our first hardware product, built by Invisible Computers, handcrafted in Germany 🇩🇪
Cal Ink (7.5") and Cal Ink Pro (10.2")
- no color, just e-ink 🔅
- no operating system, just always on 🖼️
- no AI, but magical 📆
Win a signed copy by retweeting and commenting "What do you use @calcom for?"
or preorder:
→ shop.invisible-computers.com…
→ shop.invisible-computers.com…
Michael Bang retweeted
any database is just a WAL with indexes
then it’s down to tradeoffs in guarantees, storage and indexes
Need to explain reinforcement learning to non-technical people? Show them this video: youtube.com/watch?v=NUl6Qikj…
It uses the racing game Trackmania to give an *excellent* intuition for how it works.
Such a great video @yoshtm1!
Michael Bang retweeted
Negative space in design is one of the most important aspects that you need to take care of. TigerBeetle talks about this explicitly in their TigerStyle document -
"The golden rule of assertions is to assert the positive space that you do expect AND to assert the negative space that you do not expect because where data moves across the valid/invalid boundary between these spaces is where interesting bugs are often found. This is also why tests must test exhaustively, not only with valid data but also with invalid data, and as valid data becomes invalid."
This focus on negative space also resonates with the Design By Contract paradigm first coined by Bertrand Meyer in connection with the design of the Eiffel programming language.
Similar thoughts are also expressed in Floyd-Hoare logic, with a set of logical rules for reasoning rigorously about the correctness of computer programs. The idea is to assert negative paths equally exhaustively as the positive ones.
And the most important part of using this assertion based programming is to have them in the production code as well.
